Glossary

A

Actor - Swift concurrency primitive for thread-safe state management

API - Application Programming Interface

D

Deepfake - Synthetic media created using AI to manipulate visual/audio content

DAN - “Do Anything Now” - ChatGPT jailbreak technique

G

GAN - Generative Adversarial Network - AI architecture for generating synthetic content

J

Jailbreak - Technique to bypass AI safety restrictions

P

PII - Personally Identifiable Information

Prompt Injection - Security vulnerability where malicious input manipulates AI systems

S

Sanitization - Process of removing dangerous patterns from input

System Prompt - Instructions that define AI behavior (should never be exposed)

T

Threat Score - Numerical assessment of input danger level (0-1 scale)